In this study we investigate how both local environment and individual characteristics explain incidence of corruption. More specifically, we explore how city size, and residing in a capital city influence the two aspects of corruption, notably in individuals' contact with officials, and in the use of courts. Building upon Storper's argument (2013) we further analyse the effects of local institutional trust (bridging), and community embedded trust (bonding), on incidence of corruption. Finally, we also investigate how access to elements of information and communication technology, individually and spatially defined (within the local social neighbourhood), affect corruption. To test our hypotheses, we apply a multilevel Heckman selection probit model to European Bank for Reconstruction and Development survey data. The sample covers over 26,000 of individuals in 35 countries in 2010. While our results suggest that categories of some individuals, business owners in particular, are more likely to face corruption, and therefore need more policy attention, we also find that the differences in environment play a critical role. More specifically our results suggest that larger cities are more prone to officials' corruption than medium and small ones. However, we also show that capital cities are different from larger cities in that they seem to exhibit lower corruption levels for both, courts and officials. We interpret the latter association as related to the structure of social and political connections. Larger cities are often more fragmented than capital cities in terms of power. Larger cities with many small jurisdictions imply localities where consistent expectations are easier to achieve, so individuals more likely to adopt patterns that other individuals practice, including corruptive behaviour, being trapped in a circle of corruption, where corruption becomes a (local) social norm. At the same time, capital cities are typically less fragmented and more centralised in terms of power, having metropolitan governance structures; they have bigger, more internally heterogenous jurisdictions. Importantly, there is less scope for local social process of learning from other individuals to establish corruption as a local norm. We further show that the effect of the size of the city on corruption is mitigated by higher level of local institutional trust (bridging), and inbound trust proxied by trust in friends and acquaintances (bonding). The effect of the former is weaker compared to the effect of the latter. Where bonding and bridging are both present, this reinforces their moderating effect on the impact of city size on corruption. Finally, our results also suggest that in the neighbourhoods where on average individuals have higher access to elements of information and communication technology, corruption of both officials and courts is significantly lower.
